Recent years have seen increasing promises of using antibodies as diagnostic and therapeutic agents for various disorders and diseases. Many research and clinical applications require large quantities of functional antibodies, thus calling for large scale, economic production systems to be employed. Particularly useful is the recombinant production of antibodies using a variety of expression hosts, ranging from prokaryotes such as E. coli or B. subtilis, to yeast, plants, insect cells and mammalian cells. Kipriyanov and Little, Mol. Biotech. (1999), 12:173-201.
Compared to other antibody production systems, bacteria, particularly E. coli, provides many unique advantages. The raw materials used (i.e. bacterial cells) are inexpensive and easy to grow, therefore reducing the cost of products. Shorter generation time and ease of scaling up make bacterial fermentation a more practical means for large-scale protein production. The genomic structure and biological activity of many bacterial species, such as E. coli, have been well-studied and a wide range of expression vectors are available, making expression of a desirable antibody more convenient. Compared with eukaryotes, fewer steps are involved in the production process, including the manipulation of recombinant genes, stable transformation of multiple copies into the host, expression induction and characterization of the products. Pluckthun and Pack Immunotech 3:83-105 (1997). In addition, E. coli permits a unique access to random approaches. Because of the unparalleled efficiency for transformation by plasmids or transfection by phages, E. coli systems can be used for phage library construction of many types of antibody variants, which is particularly important in functional genomic studies.
Currently, bacterial systems are used primarily to produce antibody fragments. Like any other heterologous proteins, antibody fragments can be produced in E. coli either through refolding of inclusion bodies expressed in the cytoplasm, or through expression followed by secretion to the bacterial periplasm. The choice between secretion and refolding is generally guided by several considerations. Secretion is generally the faster and more commonly used strategy.
In contrast to the widespread uses of bacterial systems for expressing antibody fragments, there have been few attempts to express and recover at high yield functional intact antibodies in E. coli. Because of the complex feature and large size of an intact antibody, it is often difficult to achieve proper folding and assembly of the expressed light and heavy chain polypeptides, resulting in generally unacceptably poor yield of reconstituted tetrameric antibody. The expression of full length antibodies in the E. coli periplasm can lead to significant aggregation of the precursor chains. While it appears that heavy chains polymerize through their cysteine residues, the nature of the heavy chain aggregation is unknown.
Furthermore, since antibodies made in prokaryotes are not glycosylated, thus lacking the effector functions, the art has suggested that E. coli would not be a useful system for making intact antibodies, especially in light of the significant problems posed by unwanted precursor chain self aggregation and low yields. Pluckthun and Pack (1997) Immunotech 3:83-105; Kipriyanov and Little Mol. Biotech. 12:173-201 (1999); Pluckthun et al. (1996) in ANTIBODY ENGINEERING: A PRACTICAL APPROACH, pp 203-252 (Oxford Press); Pluckthun (1994) in HANDBOOK OF EXP. PHARMCOL vol 3: The Pharmcol. of Monoclonal Antibodies, pp 269-315 (ed. M. Rosenberg and G. P. Moore; Springer-Verlag, Berlin).
Restricting the use of prokaryotic systems merely to production of antibody fragments is unfortunate in view of the numerous advantages of prokaryotic systems over eukaryotic systems as described above. This is particularly relevant because recent developments in research and clinical studies have suggested that in many instances, intact antibodies are preferred over antibody fragments. An intact antibody containing the Fc region tends to be more resistant against degradation and clearance in vivo, thereby having longer biological half life in circulation. This feature is particularly desirable where the antibody is used as a therapeutic agent for diseases requiring sustained therapies. Furthermore, in many instances, intact antibodies deficient in effector functions are more desirable for therapeutic uses. Friend et al., Transplantation 68: 1632-1637 (1999) describe toxic effects, such as severe cytokine release syndromes, of glycosylated CD3 monoclonal antibodies when used in humans for the treatment of acute rejection episodes of organ allografts. The CD3 antibodies cause T-cell activation and cytokine release by cross-linking the T cell receptor complex as a result of FcR binding. U.S. Pat. No. 5,585,097 describes making aglycosylated CD3 antibodies by mutating certain glycosylation site residues of native CD3 antibodies. Armour et al., Eur. J. Immunol. 29:2613-2624 (1999) describe the use of non-destructive antibodies (i.e., lacking the effector functions) specific for HPA-1a-positive platelets in therapeutic applications where depletion of cells bearing the target antigen (i.e., the platelet cells) is undesirable. Thompson, et al., J. Immunol Meth 227:17-29 (1999) show that effector functions of a fully human antibody against TGFβ2 are not necessary for use in therapy of fibrotic diseases mediated by TGFβ2. Reddy, et al., J. Immunol. 164:1925-1933 (2000) describe liability of strong antibody-Fcγ receptor binding in treating autoimmune diseases; Isaacs, et al., Clin. Exp. Immunol. 106:427-433(1996) suggest that if a pure blocking effect is required in vivo, an aglycosylated monoclonal antibody variant or a mutant engineered to prevent Fc receptor binding may be better choices.
The importance of antibodies in general for diagnostic, research and therapeutic purposes is reflected in the significant amount of effort that has been expended to study, and to modify antibody sequences and structures, from those found in natural antibodies, to achieve desired characteristics. Such attempts are well established in the art. See, for example, U.S. Pat. Nos. 6,165,745; 5,854,027; WO 95/14779; WO 99/25378; Chamow et al., J. Immunol. (1994), 153:4268-4280; Merchant et al., Nature Biotech. (1998), 16:677-681; Adlersberg, Ric. Clin. Lab. (1976), 6(3):191-205. Modifications of antibody sequences, for example those of the framework, are common. In general, however, the art recognizes that certain residues perform critical roles in conferring biochemical and functional characteristics associated with antibodies, and therefore modifications of these residues must be made with care, if at all. One such group of residues is comprised of conserved cysteine residues that form intrachain and/or interchain disulfide linkages. Conservation of these cysteines, and the apparent structural role they play, suggest that their absence or modification could lead to undesirable results. Indeed, even where attempts have been made to modify these cysteines, the thought appears to be that (i) at least a portion of the function of these cysteines must be retained in order to preserve an acceptable level of antibody integrity, function and activity; or (ii) the modification(s) can be made only in the context of antibody fragments rather than full length antibodies. See, for example, U.S. Pat. Nos. 5,892,019; 5,348,876; 5,648,237; 5,677,425; WO 92/22583; WO 99/64460; Kim et al., Mol. Immunol. (1995), 32(7):467-475. Furthermore, in situations involving absence or deletion of a genetic hinge, such as described in Brekke et al. (Nature (1993), 363:628-630), a disulfide linkage is artificially introduced to compensate for loss of disulfide linkages resulting from the absence of wild type hinge cysteines.
In light of the discussion above, it is notable that an important advance in prokaryotic production of full length antibodies was recently disclosed. Simmons et al., PCT Pub. WO 02/061090.
Despite widespread efforts and some success in improving antibody function and antibody production methods, there remains a significant and serious need for improved methods of producing antibodies in forms that are useful for, for example, diagnostic and therapeutics utilities, and at yields that are pragmatic and commercially advantageous. The invention described herein addresses this need and provides other benefits.
All references cited herein, including patent applications and publications, are incorporated by reference in their entirety.